UNIVERSITY OF GENOVA
POLYTECHNIC SCHOOL
DIME
Department of Mechanical, Energy, Management
and Transportation Engineering
Ph.D. THESIS
In
MACHINE AND SYSTEMS ENGINEERING FOR ENERGY, THE
ENVIRONMENT AND TRANSPORT
Curriculum
MATHEMATICAL ENGINEERING AND SIMULATION
XXXI CICLE
INTEROPERABILITY FOR MODELING AND
SIMULATION IN MARITIME EXTENDED FRAMEWORK
Supervisor:
Chiar.
moProf. Ing. Agostino Bruzzone
I
INTEROPERABILITY FOR MODELING AND
SIMULATION IN MARITIME EXTENDED FRAMEWORK
Abstract
This thesis reports on the most relevant researches performed during the years of the Ph.D. at the Genova University and within the Simulation Team. The researches have been performed according to M&S well known recognized standards. The studies performed on interoperable simulation cover all the environments of the Extended Maritime Framework, namely Sea Surface, Underwater, Air, Coast & Land, Space and Cyber Space. The applications cover both the civil and defence domain. The aim is to demonstrate the potential of M&S applications for the Extended Maritime Framework, applied to innovative unmanned vehicles as well as to traditional assets, human personnel included. A variety of techniques and methodology have been fruitfully applied in the researches, ranging from interoperable simulation, discrete event simulation, stochastic simulation, artificial intelligence, decision support system and even human behaviour modelling.
Glossary
AADSS Asset Allocator Decision Support System AI Artificial Intelligence
ANN Artificial Neural Network ANOVA Analysis of Variance
AODCS Attitude and Orbit Determination and Control System model APSU Automatic Proactive Simulator for Unified track generation AR Augmented Reality
ARTEM Augmented Reality TErrain interoperable Module AUV Autonomous Underwater Vehicles
CAVE Cave Automatic Virtual Environment CAX Computer Assisted Exercise
CBRN Chemical, Biological, Radiological and Nuclear CDM Crisis Disaster Management
CIMIC Civil Military Cooperation COA Course of Action
CONOPS Concept of Operations COP Cooperative behaviour COP Common Operational Picture C2 Command and Control
III
DSEEP Distributed Simulation Engineering and Execution Process DSS Decision Support Systems
DSS Decision Support Systems EC European Commission
EMF Extended Maritime Framework
FASOLT Foremost Autonomous Solutions for Operations in industriaL plant FCS Future Combat Systems
FEDEP Federation Development Process FEM Finite Element Modeling
FOE Aggressive behaviour FOM Federation Object Model GBM Glioblastoma Multiforme GUI Graphical User Interface HBM Human behaviour modifiers HIL Hardware in the Loop
HIL/SIL Hardware and Software in the Loop HLA High Level Architecture
HMI Human Machine Interactions HMI Human Machine Interface
IA-CGF Intelligent Agent Computer Generated Forces ICT Information Communication Technology
IDRASS Immersive Disaster Relief and Autonomous System Simulation IED Improvised Explosive Device
IEEE Institute of Electrical and Electronics Engineers
IPHITOS Interoperable simulation of a Protection solution based on light Interceptor Tackler operating in Outer Space
JAMS2 Joint Advanced Marine Security Simulator
JESSI Joint Environment for Serious Games, Simulation and Interoperability JSON JavaScript Object Notation
KPI Key Performance Indicator LoA Level of Autonomy
LOS Line of Sight
LVC Live-Virtual-Constructive M&S Modelling and Simulation M&T Maintenance and Training
MALICIA Model of Advanced pLanner for Interoperable Computer Interactive simulation
MDCS Multi-Domain Control Station MDTB Multi-Domain Test Bed
MEL/MIL Main Event List/Main Incident List MMI Marina Militare Italiana
V MOE Measure of Effectiveness MOM Measures of Merit
MOSES Modelling Sustainable Environments through Simulation MOOS Mission Oriented Operating Suite
MOP Measure of Performances MPA Maritime Patrolling Aircraft MRS Multi Robot Systems
MS2G Modelling and interoperable Simulation Serious Game MSpE Mean Square Pure Error
NATO North Atlantic Treaty Organization NCF Non Conventional Framework NCP Non-cooperative behaviour NGO Non-governmental organization NRE Non-reactive behaviour
OBSW On Board Software for Power Systems model ONR Office of Naval Research
ORBAT Order of battle
PS Power Systems model
RANS Reynolds-averaged Navier Stokes REG Regular behaviour
ROE Rules of Engagement ROS Robot Operating System ROV Remotely Operated Vehicle RPM Recognized Maritime Picture RTI Run Time Infrastructure SCG Scaled Conjugate Gradient SIL Software In the Loop
SISMA Medical Simulator of Astronaut including treatments, analysis and sickness models
SISO Simulation Interoperability Standards Organization SLAMS Simultaneous Localization & Mapping
SME Subject Matter Experts
SO2UCI Simulation for Off Shore, On Shore & Underwater Critical Infrastructure SPIDER Simulation Practical Immersive Dynamic Environment for Reengineering SPIRALS Space Interoperable Refilling and Advanced Logistics Simulator
SSC Pacific Space and Naval Warfare Systems Centre Pacific ST_CIPROS Simulation Team Civil PROtection Simulator
ST_CRISOM Simulation Team Crisis Simulation, Organization and Management STANAG NATO Standards
TT&CS Telemetry, Tracking and Communications System UAS Unmanned Autonomous Systems
VII UAV Unmanned Aerial Vehicles UGV Unmanned Ground Vehicle UMS Unmanned Maritime Systems
US DoD United States Department of Defence USV Unmanned Surface Vehicles
UxV Unmanned multi-domain Vehicles VAED Virtual Aided Engineering & Design VV&A Verification, Validation and Accreditation WP Work Package
Indice
1 BACKGROUND ... 10
-CRITICAL INFRASTRUCTURES IN THE EXTENDED MARITIME FRAMEWORK... -10
Offshore Critical Infrastructure and Simulation ... 10
Coastal areas ... 14
-THE COMPLEX NATURE OF TODAY MISSIONS ... -16
-AUTONOMOUS VEHICLES ... -18
-SIMULATION IN THE MARITIME DOMAIN ... -25
-ARTIFICIAL INTELLIGENCE APPLICATIONS ... -29
-AUGMENTED AND VIRTUAL REALITY ... -32
-RESEARCH BACKGROUND ... -35
-Modelling and Simulation of Manned and Autonomous Systems for Critical Infrastructure Protection in Extended Maritime Framework: ... 37
Modelling and Simulation of Autonomous Systems for Space Exploration Projects ... 39
Use of Augmented and Virtual Reality to support Maintenance and Training ... 41
Decision Support System in the Maritime Environment using Artificial Neural Network .... 42
-M&S to support design and employment of Manned and Autonomous Systems in for industrial applications ... 45
Definition of M&S Methodologies ... 47
M&S Testbed for Autonomous vehicles ... 47
M&S for supporting logistics operations in NATO... 48
-IX
M&S AND PATH OPTIMIZER ... -52
-TARGETS ... -61
-INPUT FILES ... -63
-EXPERIMENTATION ... -64
-OUTCOMES ... -78
-3 CRITICAL INFRASTRUCTURE PROTECTION... 80
-CASE STUDY ... -80
-PURPOSE OF THE SIMULATION ... -82
-SCENARIO AND MODEL DESCRIPTION ... -83
-EXPERIMENTAL CAMPAIGN ... -86
Number of Successful Recognition ... 87
Average Time to Accomplish Recognition ... 90
-OUTCOMES ... -92
-4 VIRTUAL AND AUGMENTED REALITY ENABLING SERVICES ON DISTRIBUTED ASSETS ... 94
-THE PROPOSED SOLUTION ... -96
-SERVICE DISTRIBUTED SYSTEMS ... -98
Gas Container Service ... 98
-VV&T ... -102
-OUTCOMES ... -103
-5 ARTIFICIAL INTELLIGENCE APPLICATIONS IN MARITIME SCENARIOS ... 105
-CONCEPTUAL MODEL ... -105
-PURPOSE OF THE ANNEXPERIMENTATION ... -107
-NEURAL NETWORK DESCRIPTION... -110
-TESTS AND EXPERIMENTATION ... -111
-NETWORKS TRAINING PERFORMANCES ... -116
-OUTCOMES ... -118
-6 MODELLING POPULATION REACTION TO ATTACKS ON CRITICAL INFRASTRUCTURES ... 119
-MODELLING ADDRESSING DISASTERS &CRITICAL INFRASTRUCTURES ... -120
-NATOEXPECTATIONS FROM DISASTER SIMULATION ... -124
Joint Research Activities ... 125
-EXPERIMENTATION PLANNING ... -127
-FEDERATION,FEDERATES &SCENARIO ... -128
-MODELLING IMPACTS ON POPULATION ... -130
-VV&A ... -135
-GIS AND DATA FOR DECISION SUPPORT ... -138
-OUTCOMES ... -140
-7 SIMULATION OF POWER PLANT ENVIRONMENTAL IMPACTS WITHIN THE EXTENDED MARITIME FRAMEWORK... 142
-SIMULATION MODELS ... -146
-MODELING DIFFERENT POWER PLANT EMISSIONS ... -147
Models of the EI impacts on EMF Flora ... 148
Models of the EI impacts on Marine Fauna ... 149
Economic Models ... 149
Modeling Policies, Taxations, Rules and Regulations ... 150
-XI
APPLICATIVE CASE STUDY ... -153
-OUTCOMES ... -155
-8 INTELLIGENT AGENTS & INTEROPERABLE SIMULATION FOR DECISION MAKING IN JOINT OPERATIONS ... 156
-HUMAN BEHAVIOUR MODELLING ... -159
-SIMJOH_VISFEATURES ... -163
-HUMAN BEHAVIOUR MODIFIER ... -164
-INTEROPERABILITY ... -167
-SCENARIO ... -169
-EXPERIMENTAL ANALYSIS ... -170
-OUTCOMES ... -178
-9 MODELLING AND SIMULATION FOR SPACE APPLICATIONS ... 182
NANOSATELLITES FOR BIOMEDICAL EXPERIMENTATION ... 185
M&S SUPPORTING SPACE MISSIONS ... 186
NANOSATELLITES FOR SPACE EXPERIMENTATION ... 188
GENERAL ARCHITECTURE AND MODEL DESCRIPTION ... 192
DESCRIPTION OF THE DIFFERENT MODELS ... 193
OUTCOMES ... 203
10 CONCLUSIONS ... 205
BIBLIOGRAPHY ... 210
Introduction
Introduction
This thesis proposes an innovative common environment for maritime Modeling and Simulation (M&S) with the intent to support and guide the engineering processes by adoption of marine interoperable simulators.
The thesis reports on the adoption of architectures for developing simulators able to be federated by using HLA Standard (IEEE 1516 evolved) in order to address the multi domain marine context here defined as Extended Maritime Framework (EMF). Hereafter is highlighted the architecture and the modelling approach to be followed to guarantee modularity, easy development and the capability to be integrated in the EMF scenario for Joint Naval Operations; the research provides also guidelines for experimentation and Verification, Validation and Accreditation (VV&A). Furthermore the proposed architecture and modelling guidelines identify the different modular components addressing each subject/entity by defining the required objects, attributes and interactions.
By this approach the simulators of the different assets and parties could be easily integrated in this EMF Federation. In particular the thesis focuses on federations for the development of new marine persistent and flexible capabilities based on innovative autonomous systems operating within the EMF; in this document the author reports on the researches performed during the Ph.D. related to the use of federations for testing and experimenting on new Autonomous Underwater Vehicles (AUV) integrated with traditional assets.
Introduction
Marine domain is a strategic resource for world evolution, most of world population is leaving over the coastlines and the large majority of goods are shipped; in addition, oceans are holding crucial resources for human activities from energy (e.g. oil rigs, underwater pipelines) to food (e.g. fishing); ports represents very important nodes for Nations and maritime trading lines are critical for guaranteeing development. Due to these reasons protection and control of sea is a strategic issue.
Today, the dimensions affecting the oceans are extended over several domains; the author is adopting the definition of Extended Maritime Framework (EMF) as combination of: Sea Surface, Underwater, Air, Coast & Land, Space and Cyber Space; almost all activities in the marine context are interacting with entities and resources located evenly over these domains: for instance a ship is communicating through satellites, accessing web resources, interacting with coastal ports during loading/unloading operations.
This context is defined EMF and represents the mission environment to be investigated in order to develop innovative solutions. The author is currently involved in projects related to Researches in civil and defence applications within this context so it is evident the opportunity to create a simulation environment able to support investigations and virtual experimentation in EMF.
In addition, today the introduction of autonomous new systems introduces the necessity to evaluate new configurations and operational models to utilize these entities (e.g. Autonomous Underwater Vehicles or Unmanned Surface Vehicles) guaranteeing collaboration and synergy and distributing tasks and workload in the most efficient way.
Introduction
Use of autonomous systems in collaborative modes as well as in competitive scenario represents a very challenging context, especially in EMF, due to the high influence of the different domains on the capabilities of the different vehicles and platforms.
Due to these reasons it is evident the necessity to create a simulation framework to study this context; indeed, the EMF represent a very good framework for general purpose use of new technologies that could be readapted to address defence and littoral protection as well as to support oil & gas or shipping operations.
In general, the resources interoperating over EMF represent a heterogeneous network, therefore communications as well as sensor performance are highly affected by the different domains.
Therefore, the complexity of the context as well as the high number of complex interactions among the entities does not allow approaching the problem with traditional methodologies; due to these reasons the author have been requested to develop and interoperable simulation architecture devoted to be used as a test bed for new systems operating in the EMF.
Modeling and Simulation (M&S) allows conducting experimental analysis over complex systems where different elements are dynamically interacting; in particular M&S provide a cost-effective test bed and verification tool reducing potential failures in real trials. Its applicability to Hardware and Software in the loop helps engineers in finding errors in the system embedded software and to have better insights in operation and dynamics (Montelo & Furukawa 2010).
Introduction
The goal of Modeling and Simulation (M&S) within this project is to provide both test-bed capabilities at mission level to support and analyse performances of unmanned system accomplishing tasks integrated in a traditional asset, and at physical level to provide support for virtual prototyping.
Furthermore, existing standards related to this topic have been studied to contribute to their evolution and eventually to develop new ones aiming to ensure future systems to born interoperable.
A preliminary test-bed design is developed simulating a multiple vehicle scenario. In this first simulator interoperability will be augmented to connect simulator with middleware such as ROS and MOOS, and to connect to Hardware in the Loop (HIL).
It has to be said that Unmanned Maritime Systems (UMS) are still a relatively immature technology, thus the development of a Multi-Domain Test Bed (MDTB) and a Multi-Domain Control Station (MDCS) provide the capability to carry realistic and extensive simulations. The proposed simulation will be based on the High Level Architecture (HLA) approach standardized as the IEEE 1516-series after being addressed by NATO M&S Master plan (NATO M&S Master Plan); this standard architecture has been developed to provide a flexible approach to support interoperability and distributed modelling and simulation. The HLA defines an integrated approach providing a common framework for the interconnection of interacting simulations. HLA has been developed by the Simulation Interoperability Standards Organization (SISO) and published by IEEE. To date it is a prescribed or recommended standard used by NATO as well as national department of defence. Its
Introduction
structure allows tools provided by different suppliers to become interoperable, thus reducing cost, time and risk for end users.
Traditionally HLA has been developed for defence applications, for instance to train different pilots skills in flight simulators or at Command and Control (C2) level to train officers ability to take decisions. Nowadays its applicability spans from space missions, where emergency situation can be trained without risks, to joint operation between police, fire fighters, Red Cross, from Air Traffic Management, to off-shore oil production and so forth.
As a matter of technical description of the intended architecture to use, HLA topology is a Service Bus. Each system, so called Federate, has one connection to the service bus, called Run Time Infrastructure (RTI) providing information, synchronization, and coordination services. A common set of services is agreed in the Federation Object Model (FOM) containing descriptions of the information exchanged in the federation. Each federates provide-publish or consume-subscribe the particular service that it is interested in through the RTI (Moller 2006).
In HLA context, federated simulators interoperate between each other through RTI publishing/subscribing objects or objects attributes, and interaction between objects published in the current simulation. When an object is published in the federation that specific object starts to exist for other federated simulator and thus can interact with other instances generated in the simulation.
Introduction
One of the main goal of HLA is to standardize techniques for time management among heterogeneous simulators and systems included in the federation. The adoption of HLA Standard guarantees the possibility to operate both in fast time and real time; models have then to be time constrained and time regulating. This is pursued in order to run long simulations in fast time, for instance to train capability assessment, and to operate in real time to involve real hardware in the simulation ensuring proper hardware performances. A valid example of how much simulation has gained his role in engineering processes in maritime related fields is the use of simulation tools to analyse the structural behaviour such as Finite Element Modeling (FEM) codes or the fluid dynamic behaviour with Reynolds Averaged Navier Stokes (RANS), Potential Method, Panel methods, Turbulence models codes, and so forth. Such tools are nowadays widely used and would be unthinkable to carry out a brand new design without them.
Other different types of simulators are used as virtual prototypes offering the opportunity to crosscheck the design versus performance before building expensive physical prototypes. The use of interoperable simulators allows developing a simulator for each single relevant actor in the play making requirements checking more direct and addressable. As a matter of example, in an extended maritime framework (EMF) each vehicle may be a single simulator; at the same time vehicles can be made out of single simulators constituting a system. This approach suits different applications even if this research focuses especially on virtual prototyping by building up an interoperable simulation using dynamic models of the controls, sensors, physics and communications.
Introduction
The final goal of this simulation environment is currently to support the engineering of new capabilities in the EMF based on collaborative use of autonomous systems.
The proposed federations are thus suitable for other kind of applications including: training, education, capability assessment, operational support, tactical decision aid, mission rehearsal, etc.; models with different resolutions can be integrated in this architecture to address different applicative contexts along the life cycle of the new solution; for instance it is possible to support Measures of Merit (MOM), Measure of Performances (MOP) and Measure of Effectiveness (MOE) in evaluating a specific mission environments respect alternative Concept of Operations (CONOPS) or technological solutions;
Simulations in EMF takes place in scenarios made by the six spaces listed below: 1. Underwater
2. Surface (Including maritime traffic simulation) 3. Air (assets supporting land and surface operations) 4. Land/Coast (harbours and infrastructures)
5. Cyber space (the presence of different nodes such as unmanned vehicles, and other assets make this a critical space)
6. Space (satellite)
Due to the relevance of the EMF as strategic scenario to be controlled and defended, innovative assets have to ensure persistency and resiliency to guarantee a continuous service. Unmanned technology has been explicitly developed for such an issue resulting in a cost-effective asset to be integrated with traditional one to achieve global mission tasks. Patrolling
Introduction
and environment analysis are two unmanned related field. The interoperability between unmanned vehicles with a patrolling vessel extends patrolling capacity of the system lightening operational and life cycle costs, with software embedded in control stations helping officers to track and select targets and decipher malicious behaviour potentially reducing reaction time and increasing possibility of success. To ensure proper efficacy of these tools, tests have to be carried out, hence simulation role gain relevance in this context. The use of unmanned vehicles for environmental analysis extends capability to research and collect data in locations and for mission duration not perceivable by a manned or Remotely Operated Vehicle (ROV), the possibility to achieve mission duration of several continuative days, or even weeks, with reduced human contribution is precious in environmental research related fields.
Autonomous systems are growing in relevance, as much as “The US Congress has mandated that by the year 2015, one-third of ground combat vehicles will be unmanned or autonomous” (Weiss 2011). Interoperability, and hence interactions, of different type of autonomous system either between each other or with manned system is getting more and more an issue. The decision-making process can be generated as simulations to tap the brain of the machine.
For Command and Control (C2) in this research is meant the processes of information gathering, interpreting that information so as to derive a perception of the world, and making decisions on how to respond to that perception (Lakin et al. 1998).
Introduction
The number and diversity of sources involved in the information gathering process during an operation in EMF, together with data affluence rates make C2 role an increasingly complex and difficult task. In this project C2 is regarded as Software to be federated (SIL) in the simulation federation (Bruzzone et al. 2018).
This introduction highlights the versatility of simulation as a tool to provide a test-bed to conduct extensive trials of dangerous scenarios like EMF in a safe and secure context. Future efforts will be put on development of interoperability between simulator and hardware in the loop via middleware as a step ahead in the direction of a live exercise where real vehicles are used in sea trials during simulations.
As well as for Hardware, work has to be done about SIL as a matter of time management to already build software to be part of a federation.
1 Background
Critical infrastructures in the Extended Maritime Framework.
Offshore Critical Infrastructure and Simulation
Many on-shore and off-shore installations are operating in the Extended Maritime Framework and need to be addressed. Specifically, it is important to outline that off-shore installation are complex to be protected at a reasonable costs due to their distance from consistent human workforce, as happen with off-shore platforms, off-shore wind farms, underwater pipelines and cables. From this point of view, protective solutions should be activated to cover different domains. It is evident the complexity of this framework and the necessity to integrate different domains, approaches, platforms, systems and procedures within a highly stochastic environment. As matter of facts, the use of M&S represents a very good opportunity to face these challenges and to study this complex context (McLeod 1982, Banks1998, Bossomaier 2000, Waite 2001).
The research group of the author have investigated since long time the protection of critical infrastructures in marine domain and in energy sector. In this thesis, the author proposes a systemic approach devoted to integrate innovative technologies over different kind of platforms to guarantee high level of protection with low costs based on the integration of autonomous systems and AI (Artificial Intelligence).
The research proposes cases related to the protection of off-shore platforms by using autonomous systems able to identify threats through innovative procedures; to this aim, the research includes high level models of the performances of specific algorithms and sensors on autonomous platforms for face recognition of the crew of suspected boats at large distance reducing the risk of false alarms and extending protective area. In this context the use of non-lethal weapon is crucial and this approach represent a very good example to improve the protection as well as safety and reliability.
Indeed, unmanned systems could be employed to extend the range where it is possible to identify the threats, to anticipate them and to increase the time available to adopt countermeasures (Ören & Longo 2008, Bruzzone et al 2011c). As anticipated, this approach is beneficial also to reduce the false alarms, increasing the capacity to discriminate between real and false alarms enhancing protection system credibility.
Today, the innovative unmanned systems technology is often not completely autonomous, but needs to be integrated with other traditional assets (e.g. equipment devoted to be used to intercept, discourage or engage threats) often operated by humans; in addition, the unmanned systems require usually operators and the operative and engagement procedures are driven by the decision makers (Longo et al. 2014). In this research the author adopts the Modelling, Interoperable Simulation and Serious Games (MS2G) for addressing these aspects in order to create a framework that could be used in multiple ways: evaluator of capability assessment for these innovative solutions, training equipment for unmanned systems operators and
simulator for the definition of policies and procedures (Mosca et al 1996, Kuhl et al. 1999, Massei & Tremori 2010, Guo et al. 2011, Bruzzone et al. 2012).
But what is MS2G? Nowadays simulation is evolving and presents new paradigms as Modelling, Interoperable Simulation and Serious Games, which has been developed combining the concept of HLA interoperable simulation with the serious gaming approach. This new paradigm highlights the potential of employing immersive solutions, allowing users to understand and adopt M&S ensuring interoperability standards and high fidelity level of the models developed. It is furthermore necessary to underline that this approach enhances the development of modular interdisciplinary solutions that can be adopted in new application fields (Elfrey 2006, Bruzzone et al. 2014 a, 2014 b, Bruzzone 2018). As a matter of example, this objective can be achieved with the introduction of new educational paradigms to support the evolution of potential users. Considering today opportunities is paramount to foresee future challenges. In facts, complex scenarios that can be addressed by MS2G, combined with modern technologies (e.g. Data Science, AI & Machine Learning, and Internet of Things) allow the creation of innovative contexts, generating, at the same time, new challenges. Technological evolutions shall never be underestimated.
In this document it is reported the experimentation carried out to evaluate the potential to easily train not very skilled operator in conducting such scanning procedures; the experimentation carried out with unmanned systems users allowed to evaluate the effectiveness of the MS2G solution proposed to train the operators as well as to evaluate the benefits provided by augmented reality aid and other specific algorithms (i.e. face
recognition) for what can be referred as wide range detection (Raybourn 2012, Bruzzone et al. 2014).
Among critical infrastructures the ones related to energy industry are particularly significant and it is interesting to note that while technological accidents in the energy industry have been deeply investigated over the last decades, the issue of attacks on energy infrastructures is gaining increasing importance as production and transit areas are evolving into politically unstable and unreliable frameworks. It is therefore necessary to consider energy domain under the security perspective for risk assessment (Burgherr et al. 2015). Indeed, the discussion arises on how to optimize security of critical infrastructure facing budget constraints, technological innovation and new competitive threats. The fields of investigation include Patrolling, Sensor Coverage and Interference, Domain Protection and Blocking (Bruzzone et al. 2009, Megherbi & Xu 2011, Kranakis & Kriznac 2015). To this end, many actors (e.g. EC, US DoD, NATO and Academic Institutions) are investigating innovative options (e.g. Autonomous Systems, Manned Patrolling Assets) for protecting critical infrastructures against asymmetric threats in the maritime environment using multi-agent simulation and interoperable simulation (Enters et al. 2002, Smith 2002, Lucas et al. 2007, Matusitz 2013, Massei et al. 2014, Bruzzone et al. 2015a). Ongoing researches are oriented to the definition of multi-layered architectures for reconfigurable autonomous assets (Brdys 2014), as well as models to support decision making process based on innovative techniques such as Artificial Neural Network and Genetic Algorithms (Longo 2010, Bruzzone et al 2015a) and Game Theory (Ordónez et al. 2013, Vorobeychik & Letchford
2015). The diffuse employment of unmanned systems in new operative scenarios implies the necessity to design innovative training sessions for operators through Live-Virtual-Constructive (LVC) Simulation (Vince et al. 2000, Ratliff et al. 2010, Bruzzone & Longo 2013c) capable of providing rapid and efficient knowledge and skill development for unmanned systems operators (Rowe et al. 2015). This necessity is even underlined by the availability of new technological contents, such as Augmented Reality, with which operators need to interface (Miller et al. 2014). One issue is to manually pilot the unmanned aircrafts remotely by using camera image streaming and sensors information (Yang et al. 2010) in particular for complex operations such as taking off, landing or docking and/or low altitude flight. Over the years, in order to avoid catastrophic damages to assets and increase missions’ success rate, simulation-based procedures have been designed for training operators on mission specific operational scenarios in advance (Javaid et al. 2013). Often simulation for adaptive learning is adopted to improve time-critical decision-making skills (Longo 2012, Abhyankar et al. 2014). Therefore, it is of outmost importance the development of common standard in military training and computer game simulators domains to simplify development of new concepts and to increase capability to achieve common goals reducing negative crossover (Kuhl et al. 1999; Svane & Karisson 2003).
Coastal areas
Off-shore installations are not the only critical infrastructures considered in the Extended Maritime Framework. In reference to crisis scenarios it results evident an increasing trend
Bruzzone et al. 2018). As a matter of facts, the climate change, population growth together with urbanization, industrial activities and environmental impacts are creating major causes that are expected to reinforce these events in the future (Milkov 2017, Bruzzone et al. 2018a, 2018b). In particular, the urbanisation is rapidly changing the distribution of population around the world: for instance, since last decade, more than one half of the world population lives in cities and this trend is continuously increasing (UN 2014) and many of these towns are located in coastal areas resulting vulnerable to the growing impact of extreme weather conditions (Bruzzone et al.2014b, 2017c).
In addition, the climate change and urbanisation jointly influence stability and are likely to increase the pressure on public authorities to faces these challenges as well as to task military forces for operating in large urban environments in response to instability situations such as: Man-made disaster, e.g. explosions, toxic agent contamination (Bruzzone et
al.1996b, 2014c, 2015b)
Large scale natural disaster, e.g. earthquake, flood, tsunami (Bruzzone & Massei 2006; Diaz et al.2013)
Mass migration (Bruzzone et al.2017b)
Epidemics & Pandemics (Bossomaier et al. 2009)
Inner city turmoil (e.g. social unrest, riots) and armed conflicts (Ören &Longo 2008; Bruzzone et al.2011a)
This obviously suggests the opportunity to develop new Decision Support Systems (DSS) based on interoperable simulation to address these needs being able to combine multiple models (Bossomaier & Green 2000).
It should be pointed out that natural disasters generates even more negative synergies with the emerging actual geo-political scenarios where a large number of social and political instabilities are creating very critical environments (Hsiang et al. 2014). In facts the rising number of fragile states at risk of instability and civil conflicts as well as the increase of terrorist attacks are factors that make even more difficult the crisis management (Duffield 2014).
The complex nature of today missions
Nowadays, Naval Maritime Interdiction framework includes missions of complex, specific and articulated nature. Bright cases are controlling and surveying migrants’ flows, especially when accessing south Europe through sea routes, and contrasting piracy.
Performing these activities, the Navies involved need to operate effectively and efficiently through sustainable actions to reach durable results while facing limited resources and diminishing budgets.
Furthermore, the available assets are designed for different purposes and deployment of such units in a non-specific operative point imply a significant effort in term of cost and adaptability and waist of precious assets outside their key specific applications.
Another crucial element is the strong dynamism of the phenomena under analysis. In fact, the players tend to respond quickly to countermeasures while dynamically adapting their strategies, forcing the Navies to revise planning and procedures continuously.
Indeed, the actors are often aggregated in different structures, working both as a single National Navy and as part of different coalitions (e.g., EU, NATO) in presence of other entities (e.g., NGOs, other Navies).
In addition to all these elements, cooperative operations are often very intense, widely distributed over a geographical area and strongly stochastic, which makes planning more difficult. Indeed, modern Navies need to integrate Maritime Interdiction operations (among others) with innovative technological solutions, integrating autonomous systems and advanced data fusion combining different sources (e.g. public information and military coverage); the configuration and calibration of these innovative systems and their integration with traditional assets for the intended use get large benefit from Modeling & Simulation (M&S) to evaluate the overall performance and understand the specific procedures to be adopted and to avoid/anticipate potential integration problems (Massei et al. 2011).
All these issues are subjected to stochastic factors such as failures, false alarms, times, costs, and the simulation is the cornerstone to support the analysis of this innovative operational context.
As matter of facts, the systematic use of the M&S provides strategic support in decision-making process improving both the effectiveness, the flexibility and the robustness of the planning (Bruzzone et al. 2014b; Bruzzone et al. 2013e).
Autonomous Vehicles
The inclusion of autonomy within unmanned vehicles creates opportunities for new roles and activities and enable possibility to improve operational reliability (Bruzzone et al, 2013c).
There are many ways to define the different kinds of autonomous systems such as UxV, mobile robots, intelligent systems, UAS (Unmanned Autonomous Systems), etc.; in facts there are several types of autonomous systems enabling the possibility to operate over multiple domains (i.e. Air, Sea Surface, Underwater, Land, Space, Cyberspace). In recent years, Unmanned Autonomous Systems are becoming very popular and their capabilities are improving from different points of view:
Functionalities
Operational capabilities Autonomy levels
One of the biggest challenge for these systems is to address new complex operational roles involving collaborative and competitive tasks among both Machine-Machine Interaction and Human Machine Interactions. These tasks are quite hard since they involve several disciplines such as Artificial Intelligence (AI), Soft Computing, Mechatronics and the need of guaranteeing operational interoperability (Ferrandez et al. 2013; Bruzzone 2010). Many techniques could be applied in the sector of Artificial Intelligence (Swarm Intelligence Fuzzy Logic, Genetic Algorithms, Distributed Computing, etc.) and interesting results in this sector
Zacharewicz 2008; Affenzeller et al. 2009). The use of IA-CGF (Intelligent Agent Computer Generated Forces) for collaborative and competitive assets has been extensively used with very valuable achievements in different applications (Bruzzone 2010; Bruzzone & Massei 2010; Bruzzone et al.2011); from this point of view it is very important to adopt simulation interoperability standards in the virtual scenario for testing prototypes (Zini, 2012).
Usually, the main challenge for the Autonomous Systems is strongly related to the overall complexity resulting from the combination of the mission needs and the environment characteristics; in Figure 1 several existing solutions are represented according to the combination of these two aspects (Bruzzone et al. 2013b). In facts, Unmanned Autonomous Systems are evolving along years and introduce also new aspects related to the need to introduce the concept of Multi Robot Systems (MRS), that represent the coordination and cooperation among multiple autonomous vehicles (Gerkey et al. 2003). MRS are identified as a mainstream element in the future of military and security operations:
Localization of chemical or radioactive sources Target assignments
Autonomous driving in dangerous areas Perimeter control
Surveillance
Figure 1: Evolution of Complexity in Mission and Environment among UxV different Projects.
A main classification criterion for all kind of vehicles is based on the Mode of Operation; in general, the simplest mode of operation is with a “manned” vehicle that is 100% controlled by a human on board. Obviously, this category is not applicable to UxV in general as well as to UGV systems, since they are, by definition, “Unmanned”, i.e. vehicles without the presence of human operator on-board. Another class to be considered is the “tethered vehicle”: this kind of vehicle is 100% controlled by the human operator, but remotely; obviously in this case operating in Line-of-Sight is very common; when such condition is not satisfied, the teleoperation is activated; in this case the full control stays on the human driver, but the guidance is performed by using sensors or cameras, without the Line-of-Sight
A step towards more advance solutions is in the direction of autonomous or semiautonomous systems. For instance, semi-autonomous systems have at least one autonomously controlled function; it is therefore evident that the level of autonomy can vary case by case. When autonomy reaches 100%, the mode of control does not demand the presence of a human operator and the mission could be programmed in advance assigning tasks and goals. In the following figure a classification approach for autonomy is proposed.
Figure 2: Example of Different Behavioural Modes related to the Concept of Autonomy
In facts the concept of autonomy corresponds to different issues and it is possible to define different kinds of Autonomous Behaviours; indeed autonomous behaviours represent all the actions that are undertaken autonomously by Unmanned Vehicles; for instance Autopilot functions include among the others (Stodola & Mazal 2016):
Guidance Navigation Control
SLAMS (Simultaneous Localization & Mapping) Actions to Achieve a Goal
Prioritizing Tasks
Activation of Tasks based on Situation Awareness and Threat Assessment
It is interesting to propose the classification system based on the Level of Autonomy (LoA) used by National Research Council and proposed by US DARPA (Defence Advanced Research Projects Agency) using four Levels of Autonomy (NRC 2005):
Level 1: Manual Operation Level 2: Management by consent Level 3: Management by Exception Level 4: Fully Autonomous
Other classification systems with a higher level of resolutions are existing as presented in the following table (O’Donell 2003; Deyst, & Egan, 2005); in this case the classification of Autonomy level is moving from level 1 up to level 10. The simplest way is the Level 1 that consist on remote control without any decision-making capability. In this case the Line-Of-Sight (LOS) is required. When the automation level rise, the Observation Perception and decision-making ability level increase up to the level 10 in US Army Scale for Future Combat Systems (FCS), where the system is fully autonomous and totally independent without any supervision (Huang et al. 2005).
Table 1: Levels of Autonomy in the US Army Scale for the Future Combat System
Level Level Description
Observation Perception and Situation Awareness
Decision-Making Ability
Capability Example
1 Remote control Driving sensors None Remote operator steering commands
Basic teleoperation
2 Remote control with vehicle state knowledge
Local pose Reporting of basic health and state of vehicle
Remote operator steering commands, using vehicle state knowledge
Teleoperation with operator knowledge of vehicle pose situation awareness
3 External preplanned mission
World model database— basic perception Autonomous Navigation System (ANS)-commanded steering based on externally planned path
Basic path following, with operator help
Close path following intelligent teleoperation 4 Knowledge of local and planned path environment
Perception sensor suite Local plan/replan— world model correlation with local perception
Robust leader-follower with operator help Remote path following—convoying 5 Hazard avoidance or negotiation Local perception correlated with world model database
Path planning based on hazard estimation
Basic open and rolling semiautonomous
navigation, with significant operator intervention
Basic open and rolling terrain 6 Object detection, recognition, avoidance or negotiation
Local perception and world model database
Planning and negotiation of complex terrain and objects
Open, rolling terrain with obstacle negotiation, limited mobility speed, with some operator help
Robust, open, rolling terrain with obstacle negotiation
7 Fusion of local sensors and data
Local sensor fusion Robust planning and negotiation of complex terrain, environmental conditions, hazards, and objects
Complex terrain with obstacle negotiation, limited mobility speed, and some operator help
Basic complex terrain
8 Cooperative operations
Data fusion of similar data among cooperative vehicles (such as UAVs)
Advanced decisions based on shared data from other similar vehicles
Robust, complex terrain with full mobility and speed. Autonomous coordinated group accomplishments of ANS goals with supervision
Robust, coordinated ANS operations in complex terrain
9 Collaborative operations
Fusion of ANS and reconnaissance, surveillance, and target acquisition (RSTA) information among operational-force UGVs Collaborative reasoning, planning, and execution Accomplishment of mission objectives through collaborative planning and execution, with operator oversight
Autonomous mission accomplishment with differing individual goals and little supervision
10 Full autonomy Data fusion from all participating battlefield assets Total independence to plan and implement to meet defined objectives Accomplishment of mission objectives through collaborative planning and execution, with operator oversight
Fully autonomous mission
accomplishment with no supervision
In general, increasing the autonomy in UxV requires introduction of more advanced Controls Systems, but allow to assign high level orders.
The advance in technological sector related to sensing, controlling and computing allows to bridge the gaps between manually operated assets and fully autonomous ones (Rohde et al. 2008); this is even underlined by the creation of standard communication protocols for
remote operations (Hazra et al. 2013). The use of Augmented Reality (AR), overlaying computer graphics content to real world visual data, enhances operator situational awareness in these frameworks and his capability to control vehicles, pushing the intuitiveness of User Interface within the C2 station (Vozar & Tilbury 2011; Longo et al. 2012; Bruzzone et al. 2016c).
Simulation in the maritime domain
In 2009 the USA Office of Naval Research commissioned the Unmanned Systems Branch of the Space and Naval Warfare Systems Centre Pacific (SSC Pacific) to conduct a detailed survey and analysis of current and developing robotic technologies.
The advent of technological innovative solutions such as autonomous systems increased the flexibility of modern Navies during recent years, and while in other sectors they are already very integrated, in this context they are often not a complete operative part of the missions. In facts the capability of autonomous assets could enhance mission success and is necessary to analyse how such innovative assets should be used to finalize parameters as well as to define operational procedures and policies (Bruzzone et al. 2014a, Kaymal 2016).
In the Naval sector, both civilian and military, numerous studies have been conducted in the field of Modeling and Simulation to evaluate the benefits of innovative solutions (Bruzzone et al. 2013d); indeed the marine environment is often very conservative due to the challenges provided by the sea that requires very high reliability in an hostile framework all around the
clock and in all weather conditions; for these reasons the use of simulation to address a priori potential problems is very useful and provides precious insight in advance.
Figure 3 MALICIA GUI
Simulation of maritime operations is thus an advanced and consolidated tool, using sophisticated models to reproduce the actual behaviour of different types of vehicle involved in these scenarios. Among others, specific models of MPA (Maritime Patrolling Aircraft) and UAVs have been developed to study the advantages and possible applications (Brennan & Denton, 2004, Pereira et al. 2009). Indeed, the scenario under analysis is very complex
have a very significant influence; due to these reasons, the simulation is used for planning operations and as support in the decision-making process (Bruzzone et al. 2011b). Indeed, it is fundamental to integrate in the simulation the available models able to reproduce the weather and oceanographic conditions and their impact on the assets and vehicles (Pautet et al. 2005; Delbalzo & Leclerc, 2011; Lundquist 2013). The inclusion of environmental condition effects is beneficial as the scenario analysed consists even of small-medium size boats (e.g. pirates, fisherman) and assets (e.g. UAV, AUV, USV) greatly influenced by weather conditions (Bruzzone et al. 2011c, Slootmaker et al. 2013). Indeed, in specific weather conditions even sophisticated sensors have reduced capability to detect specific kind of targets such as small medium boats (e.g. RHIB) so it is evident the importance to model all of these aspects.
In Maritime Interdiction applications, stochastic dynamic simulation offers many advantages compared to conventional studies, which would require to simplify the problem with limited generalization capabilities. These kinds of models were developed originally due to piracy issues, having a great impact on commercial maritime traffic, estimated between 1 and 16 billion US$ by United Nations Conference on Trade and Development. This leads commanders and policymakers to task scientists to develop decision support tools generating patterns and behaviour of maritime piracy actor (Varol & Gunal 2015). Planning effective activities is even a matter of considering the trade-off between operations costs and security; such tools are based on agent-based models and game theory (Bruzzone et al. 2009, Jakob et al 2012, Jeong & Khouja 2013, Marchione et al. 2014). In fact, such a scenario is very
complex, with different types of vehicles, as surface vessels, underwater vessels, MPA, UAVs, helicopters etc. with different technical characteristics, rules of engagement and differently influenced by boundary and environmental conditions. Simulation re-creates all of these different models interacting with each other by means of IA-CGF (Intelligence Agent Computer Generated Force) (An et al. 2012, Bruzzone 2012, 2013b, 2015a). Intelligent Agents recreate the behaviour of pirates present in the simulation reproducing rational and emotional reactions to patrolling assets (Bruzzone et al. 2015b, Bruzzone 2017) and the autonomous assets follow an action/reaction logic that allow them to interact with each other, thus creating a dynamic simulation (Bruzzone 2015c).
Considering the huge amount of data, interactions and information to be processed in naval operations to evaluate the effectiveness of a certain planning or strategy, the simulation is a very valuable tool (Bruzzone et al. 2011a, Cavallaro et al. 2007). By this approach, the simulator could be able to manage and use effectively unmatched input parameters and to address the uncertainty of the scenario, and within minutes it is possible simulate days, weeks or even months of activity, allowing making multiple replications of the same scene by varying the boundary conditions to understand their influence (Bruzzone et al. 2017a). Currently these models, originally developed for piracy, are under adaptation to be used in sea border protection and anti-immigration operations.
Artificial Intelligence applications
Neural Network in the maritime domains have been studied to address a certain number of different application field, ranging from development of artificial intelligence for unmanned vehicle (Zhao et al. 2016), Maritime traffic prediction (Wang et al. 2013; Daranda 2016), Behavioural Classification and prediction (Dabrowksi & de Villiers 2015; Zissis et al. 2015), radar target detection (Zhu et al. 1995), automatic damage assessment (Rose-Perhrson et al. 2000) and even for protection of marine mammals against ship strikes (Jian-Hao 2011). Automatic Target Recognition tools and methods are widely investigated in different industrial and military fields (Roth 1990, Zhao & Principe 2001). In facts the complexity of this demanding task is to be found in different aspects, such as the necessity of being time responsive, highly accurate and capable to adapt to very noisy environment. The interest in Neural Networks to address this problem is justified by the good adaptability to those constraints and, moreover, due to Artificial Neural Network (ANN) learning capabilities. Target recognition and behaviour analysis using ANN is a hot topic since late ‘90s, when first experimentation have been carried out. As nowadays, Neural Network were then developed given their applicability in scenarios characterized by high level of noise and changeability, such as moving target recognition in seaport radar images (Zhu et al. 1995; Pasquariello et al. 1998). These tools have been developed to be applied even to real satellite, radar or sonar imageries solving sensor fusion problems (Carpenter & Streilein 1998; Paes & Medeiros 2012) and have been recently applied to real time, real world maritime domain facing the hurdles not present in experimental situation such as degradation, limited collected
examples and great variety in format, influence of weather condition and volume of high resolution imageries (Tang et al. 2015; Verbancsics & Harguess 2015). Some branch of the research world is devoted to comprehend how to limit the number of images needed to properly train a neural network for maritime application, overcoming one of the biggest nowadays issues (Rainey et al. 2016).
The interest of the navies in ANN applications is highlighted by the numbers of research supported in this field by the US Office of Naval Research (ONR) founding six Navy laboratories and several universities and commercial companies to produce many of the processing tools applied in real world engineering problems and further to understand of biological and electronic neural systems investigating both biological and information processes (Miller et al. 1992). Among the ONR research, the most valuable achievements are the Classification networks developed by Dr. Lynch research team (Ambros-Ingerson et al. 1990), development of artificial intelligence for robots ad autonomous vehicles (Saillant et al. 1993) and indeed the contribution to the creation of a signal classification neural network tool developed by the Nobel Prize winner Leon Cooper now used by different DoD contractors as well as applied for credit risk assessment (Bienstock et al. 1982; Cooper & Scofield 1988).
Recent achievements in the naval domain are devoted to enhance situation awareness with data fusion approach for automated remote sensing imagery analysis applied to different sensor types, including Synthetic Aperture Radar (SAR) (Bachmann & Bettenhausen 2002) and Aerial and satellite imagery (Rerkngamsanga et al. 2016), to generate a common
operating picture providing a sound grounding for timely and reliable decision making thus increasing safety, security and readiness. Furthermore, some application is devoted to enhance on-board situational awareness, which nowadays is critical due to the increase of ship autonomy grade and consequent on-board personnel reduction, monitoring ship performances and merging signals from acoustic sensors, video image detection algorithms, integrated spectral sensors providing real time events identifications as fires, gas leaks, structural damages, pipe breaks (Minor et al. 2007).
Behavioural prediction might be expressed in different forms, depending on the final aim of the Artificial Neural Network tool. Indeed, there is a diversification in the applicability of those studies. Accordingly, behaviour might be regarded as a set of physical information about the vessel, as in the case of the ANN developed by Zissis in 2015 and Perera in 2012 in which position speed and course are forecasted using massive amount of data coming from ships Automatic Identification System (AIS). This enables analysis and supervision of single as well as classes of vessels (Bomberger et al. 2006), and used to assist port scheduling. On the other hand, abnormal behaviours may be regarded against probability of having malicious intention (Dabrowski & de Villiers 2015) classifying suspect pirates fusing information including vessels/boats track, contextual elements influencing the behaviour (e.g. time of the day, ocean condition, weather condition, season of year, location) and then feeding a classification network initialized with probabilistic relationship between variables/output. These applications are based on Dynamic Bayesian Networks showing the
feasibility of behavioural identification through tracked data and trajectory analysis based on AIS data (Castaldo et al. 2014; Mazzarella et al. 2014).
Anomalies and suspect behaviours are to date analysed with rule-based tools based on SMEs knowledge and experience, but some steps have been done following the approach of “learning from data” and the consequent use of neural network (Will et al. 2011). Despite the relevance of the phenomena there are few researches ongoing on the use of ANN applied to anti-piracy scenario and even less on maritime illegal immigration detection (Lopez-Risueno et al. 2003; Teutsch & Krüger 2010).
The author research team members have been active along the years on the research for applications adopting ANN for both industrial and military sectors (Bruzzone et al. 1998a & b; Bruzzone 2002); in particular Data Fusion systems have been integrated with ANN in Capricorn and IA-CGF (Bruzzone & Frydman 2007), while applications for identifying behaviours and/or conduct situation awareness have been used in different cases (Bruzzone & Bocca 2008; Bruzzone et al. 2010).
Augmented and Virtual Reality
The use of simulation for supporting education programs is a much consolidated approach (Ferrington et al.1992; Mosca et. Al. 1995) and even the use of Virtual Reality has been extensively applied in several sectors for training (procedures and equipment for remote operators) (Psotka 1995; Mosca et al. 1996, 1997; Moline 1997; Wilson et al 1997; Stone 2001).
Indeed, once innovative virtual frameworks are created, it becomes possible to use them also to improve maintenance and service in different ways, such as optimization of the service, logistics as well as preventive analysis on the context (De Sa et al. 1999; Bluemel et al.2003). Indeed, the innovative use of Simulation to support service of distributed systems has been effectively applied to aerospace and energy industries (Bruzzone & Simeoni 2002; Haritos 2005).
Augmented and Virtual Reality could support many aspects, with servicing and maintaining among the most promising since the introduction of these techniques (Azuma 1997). The use of mobile technologies is very interesting to address distributed systems; indeed the mobile solutions for simulation based training in external logistics have been demonstrated very successful in several contexts confirming the potential of mobile training concept (Bruzzone et al. 2004a, 2004b, 2017b; Monahan et al. 2008; Ally 2009; Lee 2011). Indeed, the interactive Virtual Worlds are effectively enabling new opportunities in training applied to procedures and operations in many sectors (Bruzzone 2009; Raybourn 2014).
In this thesis it is proposed to investigate the use of web technologies and cloud approach to support services in conjunction with AR, VR and M&S. For instance, applications to industrial plants and components based on web technologies and virtual reality are popular from almost two decades and are continuously evolving (Bruzzone 1999; Monahan et al. 2008). Even if these concepts have been investigated since many years, the recent evolution in web technologies is enabling new opportunities in service and maintenance (Bruzzone et al. 1999; Vora et al. 2002), providing the remote users with devices able to evaluate the status of the operator in order to optimize the effectiveness of training and of operation. In this sense, the use of innovative solutions based on the capability to capture physiological parameters remotely (e.g. EEG, muscular tone, cardio frequency) has a great potential in supporting operator and user supervision; from this point of view these researches are very consolidated (Orlansky 1994; Brookings 1996) and are leading to new solutions (De Crescenzio et al.2011; Bruzzone et al. 2016a).
Figure 4 SPIDER: the Virtual Immersive Interoperable Interactive CAVE for Virtual Training and Engineering
Research Background
During the first year as PhD. Student, tutored by prof. Agostino G. Bruzzone, in order to form a consistent background knowledge, the author focused on Modelling and Simulation related to appliances on Autonomous Systems, Artificial Intelligence and Intelligent Agents. In particular the topics of the researches, in accordance with the purposes identified in the research project, are M&S as a tool to design and support the employment of Manned and Autonomous Systems both for Critical Infrastructure Protection in Extended Maritime
Framework and in space environment. In addition, a branch of the research has been devoted to application of Virtual and Augmented Reality for maintenance and training.
During the second year, the topics of the researches in accordance with the purposes identified in the Research Project have been M&S tools for Decision Support Systems (DSS) in the maritime environment using Artificial Neural Network (ANN). In addition, a branch of the research has been devoted to support both the design and the employment of Manned and Autonomous Systems for industrial applications.
The author had the chance of supervising the thesis of two international students of the MIPET master, on a feasibility study, based on M&S, for the storage of CO2 into the marine environment. The study has been carried out in conjunction with the ERG Power Plant located in Priolo, Sicily, Italy.
The author has been involved in the Visiting Research Program (VRP) of NATO STO Centre for Maritime Research and Experimentation (CMRE) in La Spezia, with 1 year scholarship founded by the CMRE, starting August 1st 2017 ending August 31st, 2018.
Afterwards, the author continued to work within the CMRE as Staff Member as Modelling and Simulation Scientist.
The topic of the research focuses on the design of a simulated testbed for unmanned systems, the implementation of standard procedures and techniques to Design, Verify, Validate and Accredit distributed interoperable simulations and data collection for Decision Support System inside the alliance.
The research activity has been coupled with active participation to NATO NMSG-139 (NATO Modelling and Simulation Group) till its conclusion in September 2017 dealing with the development of innovative techniques to evaluate M&S (Modelling and Simulation) Use Risk. These participation have incliuded the participation to a course at Johns Hopkins University, Applied Physics Laboratories on VV&A and risk computation; the course has been finalized by an exam.
The research activity has been coupled with active participation to NATO MSG-147 dealing with the development of interoperable tools for Civilian-Military cooperation in Crisis Management and Disaster Relief (CMDR).
The following main topic have been addressed in the research activities:
Modelling and Simulation of Manned and Autonomous Systems for Critical
Infrastructure Protection in Extended Maritime Framework:
Critical Infrastructure Protection is a rising issues in today world; considering that most of the population lives on coastal area it is not surprising the fact that several of these infrastructures are located within marine scenario. Ports, piping, cables, off-shore and coastal on shore plants are being more and more targeted by asymmetric threats. Employing Autonomous Assets allows to drastically reduce the protection costs but requires to design new solutions. The research carried out along this first year addresses this issue with special attention to off-shore platforms respecting the opportunity to improve threat assessment by innovative solutions. An MS2G (Modelling, Interoperable Simulation Serious Game) Agent Driven stochastic simulation for reproducing a combined used of autonomous and traditional
assets has been developed to identify threats as well as the possibility to use it for training Unmanned Aerial Vehicles (UAV) pilots. As the research goal is to investigate the potential of innovative technologies the models integrate AI algorithms for face recognitions with sensors mounted on rotary wings vehicles as support for protecting offshore platforms. In this case the simulation is mainly devoted to understand the operative advantage of a rotary wing drone employed on off-shore platform for the use of the decision maker, to provide a useful tool for the definition of design requirement of the system and to provide a test-bed to train drones operator performing recognition activities in a hostile environment facing unconventional targets.
One of the simulator proposed for this case study is SO2UCI (Simulation for Off Shore, On Shore & Underwater Critical Infrastructure) and it has been developed by the Simulation Team; SO2UCI is a simulation able to support system requirement definition phase and training on protecting Off-Shore Platforms (e.g. oil rig, gas rig), On-Shore Critical Infrastructures (e.g. ports, power plants, refineries, desalinations plants) and Underwater Critical Infrastructures (e.g. cables, pipelines) from Asymmetric Threats using conventional assets and autonomous systems. The results of the experimental campaign obtained on a test population of unskilled operators have been presented to evaluate the possibility to diffuse the use of such approach without requiring very highly qualified expertise. The experimentation carried out with unmanned aerial systems allowed to evaluate the effectiveness of the models proposed to evaluate the benefits provided by augmented reality aid and other specific algorithms such as face recognition.
Figure 5 A frame of a simulation depicting a typical Critical Infrastructure Protection scenario.
Modelling and Simulation of Autonomous Systems for Space Exploration
Projects
Space exploration is the ongoing discovery and exploration of celestial structures by the continuous improving of space technologies and future exploration missions are expected to establish a solid partnership between human crew and robots. NASA is already employing robot, tele-robots (remotely operated) and Autonomous Systems and have identified critical elements in:
Technologies enabling Autonomy Capability
Technologies able to exceed human performance (e.g. sensing, piloting, driving, manipulating, rendezvous and docking).
Technologies supporting cooperative behaviours among the autonomous systems and humans
